GPT-6.2 Astra and GPT-6.2 Sol on BenchGecko · price and context
OpenAI's GPT-6.2 Astra lists at $8 input and $40 output per million tokens, 20% below GPT-6 Astra, while GPT-6.2 Sol keeps $2 and $10 · both with a 1,050,000 token context.
What was added
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026. Both are tracked on BenchGecko together with their batch rows and the Pro variants listed on OpenRouter. The API ids are gpt-6.2-astra and gpt-6.2-sol. OpenAI positions Astra as its frontier model for the hardest projects and Sol as the efficient model for everyday work.
Context and limits
Both models accept text and image input and return text. OpenAI lists a 1,050,000 token context window, a maximum input of 922,000 tokens and a maximum output of 128,000 tokens for each. The knowledge cutoff is May 18, 2026 for Astra and June 1, 2026 for Sol. Both support reasoning tokens, and the API rejects requests that set reasoning effort to none.
Price on BenchGecko
- GPT-6.2 Astra · $8 input · $40 output · $0.40 cached input per million tokens
- GPT-6.2 Sol · $2 input · $10 output · $0.10 cached input per million tokens
- GPT-6.2 Astra (batch) · $4 input · $20 output
- GPT-6.2 Sol (batch) · $1 input · $5 output
- GPT-6.2 Astra Pro and GPT-6.2 Sol Pro · same token prices as the base models, with pro reasoning mode spending more tokens per answer
OpenAI also lists 30 minute cache writes at $10 for Astra and $2.50 for Sol. Requests above 272K input tokens are billed at a long context rate for the full request: $16 input and $60 output for Astra, $4 input and $15 output for Sol.
How it compares on price
GPT-6 Astra lists at $10 input, $50 output and $1 cached input on BenchGecko, so GPT-6.2 Astra is 20% cheaper on input and output and 60% cheaper on cached input. GPT-6.2 Sol keeps the price of GPT-6.1 Sol: $2 input, $10 output and $0.10 cached input. All four share the same 1,050,000 token context window and 128,000 token output limit, and Astra still costs 4 times as much as Sol per token.
Vendor benchmarks
OpenAI reports these figures in its announcement:
- Agents' Last Exam · GPT-6.2 Astra scores 5.2 points above GPT-6 Astra
- Terminal-Bench-Science 0.1 · GPT-6.2 Astra scores 9 percentage points above GPT-6 Astra
- GDPval v2 · Astra up 6.2 percentage points on GPT-6 Astra with 22% fewer output tokens · Sol up 5.3 points on GPT-6.1 Sol
- AA-AnalystAgent · Astra answers 62.5% of questions correctly on all five attempts
- FrontierCode 2.0 · Sol comes within 0.5 points of Astra, both at max reasoning
These are self reported results. BenchGecko has no independent scores for either model yet and will add them to the model pages as publishers report them. BenchGecko's launch day profile tests for Sol are in the launch report card.
Sources
Source: BenchGecko
Written automatically from BenchGecko's tracked data for this period; every number comes from the changelog, model, mindshare and pricing datasets. Checked against the source data before publishing.